Proposal for ACM Special Interest Group on Health Informatics We are a group of researchers from multiple academic and research institutions and would like to bring to your consideration the idea of starting a new ACM special interest group on Health Informatics (SIGHI). We have been collectively working on multiple aspects of the Health Informatics discipline for several years and have been noticing the emergence of substantial research and associated technologies in this field. The resulting effort lies in the confluence of several fields, including Computer Science, Engineering, Medicine, and Nursing. Considering the rising amount of interest in the area as well as the cross-disciplinary aspects, we believe this area has the potential for becoming a new ACM special interest group with both conferences and journals. As evidence of the growing research and practitioner interest in this area, we should point out that there are several existing Computer Science venues that accept Health Informatics contributions on specific topics and have been successful over the last few years. These venues include the International Conference on Pervasive Computing Technologies for Healthcare (Pervasive Health), the ACM-SAC Conference Track on Computer Applications in Health Care, and the SIGKDD conference (specifically in the area of mining medical data). A critical mass of a health informatics community has been established through these venues. A future SIGHI group can help grow ACM by unifying this critical mass and bringing together researchers and practitioners from the Computer Science community, the Engineering community, the medical community, and the nursing community. Health Informatics is a large field and has a wide range of aspects. The current Computer Science venues supported by existing ACM SIGs only cover a small portion of these aspects as a side effect. SIGHI can bring all these aspects front and center, and has little overlap with the existing ACM SIGs. In the long term, SIGHI could be better integrated with the medical community by defining mechanism to allow cross-listing of publications with PubMed/Medline. While our objective is to start SIGHI, we understand that this effort might require additional work. As a way to jumpstart this effort, we believe that it can be initiated by establishing a regular ACM (possibly jointly with IEEE and AMIA) Health Informatics Conference, possibly co-located with an existing conference such as AMIA (American Medical Informatics Association) annual symposium. Proposal for the ACM International Conference on Health Informatics Why is a new venue necessary? The volume of work on Health Informatics is quickly increasing with contributions from various groups in the Computer Science community. This type of research has been published in a scattered form in forums such as SIGCHI and SIGIR conferences, making it difficult to unify a critical mass of researchers and practitioners. Furthermore, Health Informatics work is interdisciplinary and requires deep knowledge in both Medicine and Computer Science. The fact that the community is inter-disciplinary as are the existing venues creates several problems. For example, suppose a researcher writes a paper mixing medical technologies and formal algorithmic derivations, including sophisticated Computer Science techniques as is becoming typical from cross-domain research groups. While such a paper can be sent to existing health informatics venues like the annual AMIA conference, the majority of the reviewers and audience are medical practitioners (e.g., physicians and nurses), often do not have the proper background in Computer Science formalism, and may have difficulty understanding it. This often causes the paper to be rejected and its merit underappreciated. Likewise, the paper can also be submitted to a Computer Science conference. In this case, the reverse problem usually happens, as the reviewers might not be able to fully grasp the medical implications, hence the paper may not get in due to misunderstandings and lack of technical depth from a Computer Science standpoint. Indeed, in medicine, a one page paper discovering an important medical fact for a single (could be rare) disease without any “deep,” associated techniques can still have substantial practical impact. Another important reason why a new venue is necessary is to develop new methods of learning from existing data and apply it to future interventions. Medicine is strife with articles that essentially re-invent the wheel, and several articles that are published are based on small case numbers due to either the rarity of the disease, lack of multi-institutional collaboration or lack of formal protocol driven studies. It is important to develop sophisticated mathematic techniques that can help synthesize a lot of the pre-existing data and provide clinicians with a composite index of published data, which can help them make decisions for future patients. Meta-analysis and systematic reviews, which are the current way to synthesize evidence, remain grossly inadequate in most fields and become outdated quickly with rapid and large volume updates in information. Our conference will also challenge our members to integrate the fields of medicine and computer science in this area. Another important aspect has to do with cultural research community differences. We believe that a cross-disciplinary field such as Health Informatics should be allowed to have its own set of guidelines and accepted norms for paper acceptance rates and peer evaluation. A final aspect has to do with how incipient the area is at this point. In general, science is an incremental process by building one’s own work on other people’s existing work. When a new area/topic is in its early stage, the first few papers are typically foundational, especially when building actual systems, and hence may not have very formal evaluation results. For example, the first WWW Google paper “The Anatomy of a Large-Scale Hypertextual Web Search Engine” provides only anecdotal examples showing which Web pages can be found using Google for a few keyword queries. Now as the WWW conference became much more selective, with acceptance ratio around 12%, such paper would most likely not to be accepted for lack of scientific rigor. A second example is Codd’s relational database model paper “A Relational Model of Data for Large Shared Data Banks” where no implementation was originally provided, but eventually led to Codd’s Turing award. A third example is the DNA double helix structure paper that was just a conjecture paper with 1,000 words at the time of publication in Nature. The final full validation of that conjecture was only done in early 1980’s, more than 20 years after the original paper was published. Our point is that if this kind of work cannot be published, other researchers will not become aware of emerging ideas. This can be a huge loss to the entire field, if more adventurous research is not given the proper forum for discussion and evolution. The important aspect in our view is to provide venues that will lead to fruitful and exploratory research so that everybody in the field can benefit from it and the field can flourish. Taking these considerations, the initial ACM venue we would like to start should have the following characteristics in terms of the research contributions that will eventually appear: (1) The maximum paper length can be 10 pages in the ACM conference format. Nevertheless, authors can write papers with any length, from 1 page to 10 pages. Such approach is adequate to address the multitude of research contributions that might arise. For example, a medical IT fact paper using many medical and computer science jargons can be 1-page long, whereas a paper describing a system can be 10-page long. All papers should be reviewed according to the same criteria regardless of their lengths. All papers can be categorized into two categories: short papers with at most 5-page and long papers with at most 10-page. Both kinds of papers will be treated equally (e.g., acceptance criteria/ratio). Long papers can have longer oral presentation time than short papers. We can also consider 1-page demo papers, each having 15 minutes in an oral demo presentation session. (2) To foster innovation, conjecture papers should be allowed by showing the supporting reasons and evidence behind the conjectures. In this case, the review process should be more accepting so that important foundational ideas can get out. The reviewers should look for interesting new facts, new applications, new frameworks, new lessons learned from building a medical software system, and/or new techniques. The evaluation of a new piece of work does not have to be very formal or complete. Evaluation by giving examples should be acceptable, to allow other people to follow on the work for improvement or more formal evaluations in future publications. A paper does not have be comprehensive and can focus on a single aspect, for example, just describing a new application without giving detailed implementation algorithms, just proposing a new technique for an existing application, or just giving a formal, complete study that was not done before. It will be implicit, however, that the proposed methods or suggested algorithms must have a practical clinical application, which may be nascent or developed. (3) The program committee must include many clinicians and many computer scientists, and papers should be reviewed by a cross-disciplinary team. The author of a contribution can also specify the type of preferred reviewers. The first type is mainly clinician reviewers, for papers describing new medical IT facts or new applications for clinicians. The second type is mainly Computer Science reviewers, for papers describing new algorithmic techniques and formal derivations, or new applications for ordinary people without medical background. The third type is a combination of Computer Science reviewers and clinician reviewers. Clinicians can evaluate the validity of medical facts and computer scientists can provide an evaluation of the computer science techniques. In this case, clinicians and computer scientists can together form an opinion on the contribution’s merit. (4) We acknowledge that very few people have deep understanding of all three fields of Computer Science, Medicine, and Nursing. In the early stage, we may not be able to enlist enough such people into the program committees. Consequently, many committee members will have expertise in Medicine, Nursing, or Computer Science, but not all three fields. Nevertheless, clinician reviewers cannot reject papers because they do not know the computer science jargons or math formula derivation. Computer scientists cannot reject papers because the paper talks about a niche market or just medical IT facts without deep computer science techniques. Remember a solution to a rare disease can benefit the society. All kinds of health informatics papers, whether about new medical IT facts/applications, computer science techniques, or a hybrid of both, should be within scope. Certain topics such as pure medical image processing papers and pure bioinformatics papers should go to their own existing conferences, but new topics such as system software for complex clinical studies that involve combinations of clinical, genetic, genomic, imaging, and pathology data, new applications in electronic medical records (e.g., using semantic technology, using continuous biomedical signals to trigger health alarms), consumer health search, and issues involving interoperability and data representation in healthcare delivery should definitely be in the scope of this new venue. (5) The review process must be expedient so that authors can get reviews in one or two months’ time after submission, allowing improvements to be made and new ideas to be incorporated. (6) The published papers are indexed in both ACM and PubMed/Medline, following the style that IEEE EMBC conference papers are indexed to allow for broad dissemination and the quick establishment of a critical mass of researchers. The open access requirements of NIH-funded research (http://publicaccess.nih.gov/) should be honored. (7) 30 minutes will be allocated for each paper presentation. In case that too many papers are submitted to and subsequently accepted by this venue, the conference organizers will employ one or more of the following strategies. First, increasing the number of concurrent sessions by grouping specific subareas. Second, decreasing the oral presentation time to 20 minutes. Third, accepting higher-score papers as oral presentations and additional papers as posters with substantial time for the poster session to foster interaction and discussions. All types of papers will be treated equally in the conference proceedings. In the peer review process, reviewers can see authors’ names but not vice versa. Initiator: Gang Luo, IBM T.J. Watson Research Center, email@example.com List of people supporting this effort: - Gang Luo, PhD. IBM Research, Gang Luo/Watson/IBM@IBMUS - Henrique Andrade, PhD. IBM Research, Henrique Andrade/Watson/IBM - Olivier Verscheure, PhD. IBM Research, Olivier Verscheure/Watson/IBM - Joel H. Saltz, MD, PhD. Director of the Center for Comprehensive Informatics, Professor in the Departments of Pathology, Biostatistics and Bioinformatics, and Mathematics and Computer Science at Emory University, Chief Medical Information Officer at Emory Healthcare, <firstname.lastname@example.org> - Blackford Middleton, MD, MPH, MSc. Director Clinical Informatics Research & Development, Partners Healthcare System, Brigham & Women’s Hospital, Harvard Medical School, AMIA Board of Directors, <bmiddleton1@PARTNERS.ORG> - Deepak S. Turaga, PhD. IBM Research, Deepak S Turaga/Watson/IBM@IBMUS - Michael D. Larsen, PhD. Associate Professor of Statistics; Faculty, Biostatistics Center, George Washington University, <email@example.com> - Daby M. Sow, PhD. IBM Research, Daby M Sow/Watson/IBM - Vagelis Hristidis, PhD. School of Computing and Information Sciences, Florida International University, <firstname.lastname@example.org> - Liangyou Chen, PhD. Research Scientist, Bioinformatics Cell, TATRC, U.S. Army Medical Research and Materiel Command, <email@example.com> - Umit V. Catalyurek, PhD. Associate Professor, Department of Biomedical Informatics, Department of Electrical & Computer Engineering, The Ohio State University, <firstname.lastname@example.org> - Kiran Turaga, MD, MPH. Surgical Fellow and Clinical Instructor, H. Lee Moffitt Cancer Center and Research Institute, email@example.com - Jinbo Bi, Ph.D. Mass General Hospital, Department of Emergency, Research Scientist, firstname.lastname@example.org, email@example.com - Yuan An, Ph.D. Assistant Professor, College of Information Science and Technology, Drexel University, <firstname.lastname@example.org> - Chunqiang Tang, PhD. IBM Research, Chunqiang Tang/Watson/IBM - Cong Yu, PhD. Yahoo! Research, <email@example.com> - Jan Claassen, MD, PhD. Assist Prof of Clin Neurology, Dept of Neurology, Columbia University, <firstname.lastname@example.org> - J. Michael Schmidt, Ph.D. Assistant Professor of Clinical Neuropsychology in Neurology, Informatics Director, Neurological Intensive Care Unit, Critical Care Neuromonitoring, Columbia University College of Physicians and Surgeons, 177 Fort Washington, Milstein Hospital, <email@example.com> - Desney Tan, PhD. Microsoft Research, <firstname.lastname@example.org> - Gregory D. Abowd, PhD. Interim Director of the Health Systems Institute. Professor in the College of Computing at Georgia Tech, <email@example.com> - Gillian Hayes, PhD. Assistant Professor in Department of Informatics, University of California, Irvine. firstname.lastname@example.org - Michael Weiner, M.D., M.P.H. Associate Professor of Medicine, Indiana University School of Medicine, Indiana University Center for Aging Research, Regenstrief Institute, Inc. email@example.com - David A. Bader, PhD. Professor, Executive Director of High-Performance Computing, College of Computing, Georgia Institute of Technology, <firstname.lastname@example.org> - Michelle L. Rogers, PhD. Assistant Professor, The iSchool at Drexel University. <email@example.com> - Vinay S Mahajan, MD, PhD. Resident in Clinical Pathology, Department of Pathology, Brigham and Women's Hospital. firstname.lastname@example.org - David W. Bates, MD, MSc. Medical Director of Clinical and Quality Analysis, Information Systems at Partners HealthCare System, Inc., Chief of the Division of General Internal Medicine at the Brigham and Women's Hospital, Professor of Medicine at Harvard Medical School and Professor of Health Policy and Management at the Harvard School of Public Health, Board Chair of the American Medical Informatics Association (AMIA), member of the Institute of Medicine of the National Academies and External Program Leader, Patient Safety Research, World Health Organization (WHO) World Health Alliance for Patient Safety. <DBATES@PARTNERS.ORG> - Richard E. Ladner, PhD. Boeing Professor in Computer Science and Engineering, Department of Computer Science & Engineering, University of Washington, IEEE Fellow, ACM Fellow, <email@example.com> - Lu Wang, MD, MSc. Medical Informatics Fellow, Clinical Informatics Research & Development, Partners Healthcare System, Harvard-MIT, firstname.lastname@example.org, email@example.com - Amar K. Das, MD, PhD, Assistant Professor of Medicine (Biomedical Informatics) and of Psychiatry and Behavioral Sciences, Stanford University, <firstname.lastname@example.org> - Tiffany Veinot, MLS, PhD, Assistant Professor, School of Information, School of Public Health, Dept. of Health Behavior & Health Education, University of Michigan, email@example.com - Joseph A. Konstan, PhD, Distinguished McKnight University Professor, Department of Computer Science and Engineering, University of Minnesota, ACM SGB Past Chair and SGB Representative to ACM Council, firstname.lastname@example.org - Andrew Sears, PhD. Constellation Professor of Information Technology and Engineering, Department of Information Systems, UMBC, email@example.com - Ashish Joshi, MD, MPH, Research Assistant Professor, Department of Information Systems, UMBC, firstname.lastname@example.org - Zhaohui Cai, MD, PhD (Pharmacology). AZ Pharmaceuticals - US, Associate Director, Informatics, Clinical Information Science, Zhaohui.Cai@astrazeneca.com - Thomas A. Finholt, Ph.D. Associate Dean and Professor, School of Information, University of Michigan, email@example.com - Kai Zheng, PhD. Assistant Professor of Health Management and Policy, School of Public Health; Assistant Professor of Information, School of Information; and Adjunct Assistant Professor of Nursing Informatics, School of Nursing, the University of Michigan. <firstname.lastname@example.org> - YingLi Tian, PhD, Associate Professor of Department of Electrical Engineering, The City College of New York and Graduate Center, City University of New York. email@example.com Partial list of topics of interest of the conference (under construction): - system software for complex clinical studies that involve combinations of clinical, genetic, genomic, imaging, and pathology data - new applications in electronic medical records (e.g., using semantic technology, using continuous biomedical signals to trigger health alarms) - consumer health search and medical information retrieval - issues involving interoperability and data representation in healthcare delivery - multimodal medical signal analysis - large-scale mining longitudinal patient records - intelligent medical devices and sensors - data management privacy and confidentiality issues - continuous monitoring and streaming technologies and analytics applied to direct and remote patient care and hospital environments - e-healthcare infrastructure design for the developing world - user-interface design issues applied to medical devices and systems - patient and institutional medical compliance automation - computational supports for patient-centered and evidence-based care - new technologies for capturing and documenting patient encounter information in electronic systems - keyword and multifaceted search over structured electronic medical records - information technologies and patient safety and healthcare outcomes - developing novel healthcare information systems for chronic disease management - Personalized predictive modeling for patient management (trauma, diabetes, sleep disorder, substance abuse etc) - Knowledge discovery for improving patient-provider communication - Better estimate of hazard functions, survival analysis - Optimization models for therapy recommendation and planning - Medical data display and visualization - Medical decision-assist systems - Medical data engineering - Physiological modeling - Assistive-living technologies - Web accessibility (Notes 1) Another important aspect has to do with cultural research community differences. We believe that a cross-disciplinary field such as Health Informatics should be allowed to have its own set of guidelines and accepted norms for paper acceptance rates and peer evaluation. For example, the AMIA conference has an acceptance ratio of around 50% and clinicians regard it as a competitive forum, whereas ACM conferences often have acceptance ratio of around 10%-15%. Note also that in many research fields such as biomedical engineering, the journal acceptance ratio is higher than 30%. In Physics, physicists in the past tended to be highly selective in reviewing research contributions so that many important papers, such as the first paper on String Theory, were rejected. Now physicists realize this problem and tend to be lenient in reviewing papers. Nowadays, a physicist can mainly send his papers to the online archive arXiv.org. If the papers are important and become widely known in the Physics community, peer recognition will naturally follow. In many cases, the relative importance of a particular research should be judged by the test-of-time, not necessarily limited by peer reviewers in the short-term conference review process. We believe that health informatics techniques and procedures elicited by research will typically fit this same long-term model and new venues in this area should be designed to accommodate it. (Notes 2) To accommodate the cross-disciplinary review process, the third type of papers will be reviewed in a non-traditional way. Each such paper is judged by three reviewers and an additional meta-reviewer. To ensure enough cross-discipline validation, every program committee member will have chance to act as meta-reviewer for some papers. The paper author will provide a cover page specifying how these three reviewers should be allocated: the number of computer science reviewers, the number of physician reviewers, the number of nurse reviewers, and which reviewer should focus on which part/section of the paper. To facilitate program committee chair’s task of assigning papers to reviewers, for each field of medicine, nursing, and computer science, every program committee member will specify in advance whether he/she has deep, superficial, or almost no knowledge in that field. Each reviewer will read the entire paper. Clinician reviewers focus more on the validity of clinical claims/applications and are allowed to explicitly admit that they do not know whether the computer science techniques are valid. Computer science reviewers focus more on the computer science techniques and are allowed to explicitly admit that they do not know how important the clinical application/implication is. After generating his/her own review, each reviewer is allowed to see the other two reviewers' reviews and adjust his/her review if needed. Then the meta- reviewer makes the final judgment on the paper, by considering the feedbacks received from computer science reviewers and clinician reviewers. The meta-reviewer may not have deep knowledge in all three fields of medicine, nursing, and computer science, but can make this up through reading reviewers’ feedbacks as well as contacting reviewers for clarification if necessary.
Pages to are hidden for
"health_informatics_SIG"Please download to view full document